FloodDiffusion 2: Efficient and Path Controllable Streaming Motion Generation

Yiyi Cai; Yuhan Wu; Kunhang Li; Tu Fangyuan; Xiangyue Zhang; Qiaoge Li; Zhixiang Wang; Kaipeng Zhang

cs.CVSep 27, 2026

FloodDiffusion 2: Efficient and Path Controllable Streaming Motion Generation

Authors: Yiyi Cai, Yuhan Wu, Kunhang Li, Tu Fangyuan, Xiangyue Zhang, Qiaoge Li, Zhixiang Wang, Kaipeng Zhang, +1 more

Organizations: Alaya Lab · The University of Tokyo · Japan Advanced Institute of Science and Technology · China Mobile Communications Company Limited Research Institute

Abstract

We present FloodDiffusion 2 (FD2), an efficient and controllable framework that builds upon FloodDiffusion (FD1), a state-of-the-art streaming motion generation model. While FD1 produces plausible motion, it suffers from low efficiency and limited controllability, as its attention design requires repeated computation over the entire history, and it lacks precise trajectory control for real-world applications. To address these limitations and improve generation quality, FD2 introduces three advances. First, Partial Attention makes finalized history representations independent of the active window, enabling KV-cached inference and shared-history packing for efficient training. Second, we establish a necessary-and-sufficient Bregman criterion for regression losses to preserve diffusion's conditional-mean velocity field. This criterion guides an FK-induced quadratic loss that incorporates motion geometry without online FK evaluation. Third, FD2 introduces precise path conditioning to control the character's root trajectory while preserving natural body motion. Experiments show that FD2 reduces training computation by 4.6 $\times$ and accelerates denoising by 11.29 $\times$ , reaching 2.303 ms per update on long sequences. Alongside these efficiency gains, FD2 improves motion quality over FD1 and achieves state-of-the-art FID scores among streaming methods, with 0.048 on SEED and 0.053 on HumanML3D.

Figures & tables

Figure 1: FloodDiffusion 2 generates continuous human motion from time-varying text prompts while following a specified path. The character transitions smoothly from walking to running and then crouched walking. Blue and red curves denote the target and generated center-of-mass trajectories, respectively; numbers indicate the generated planar speed in m/s.

Figure 2: Frame-staggered diffusion schedule retained from FD1. (a) Time-shifted denoising ramps separate finalized history, the active window, and future noise. (b) Overlapping active windows advance as frames finish denoising one by one. The illustration uses

n_{s}=4

;

\beta

denotes the noise amplitude.

Figure 3: Partial Attention enables KV caching and shared-history training. (a) History queries read their causal prefixes, while active queries read the complete history and interact bidirectionally within the active window. (b) Newly finalized frames are re-encoded causally before their keys and values are appended to the cache. (c) Isolated noisy windows share one causal clean reference, reducing repeated history computation.

Figure 4: Diffusion-compatible losses and the FK-induced quadratic objective. (a) Under the stated regularity and integrability conditions, target-first Bregman divergences with strictly convex potentials, up to a target-only term, are necessary and sufficient for uniquely recovering the conditional mean. (b) Kinematic response Gram matrices estimated from training data are averaged and normalized, then combined with an identity term to obtain a fixed positive-definite metric, avoiding online FK evaluations.

Attention	R@3 $\uparrow$	FID $\downarrow$	MM-Dist $\downarrow$	Div. $\rightarrow$
Full	$0.810$	$\mathbf{0.057}$	$2.887$	$9.579$
Causal	$0.793$	$0.096$	$3.005$	$9.369$
Partial	$\mathbf{0.814}$	$0.058$	$\mathbf{2.839}$	$\mathbf{9.438}$

Table 1: Attention-mask ablation on HumanML3D with the same 263-D configuration, checkpoint budget, and

K=1

. Diversity closer to real motion (9.503) is better.

History	GFLOPs $\downarrow$	ms $\downarrow$	GFLOPs $\downarrow$	ms $\downarrow$	MiB	$\times$
	Full, uncached		Partial, KV cached		Cache	Speedup
$0$	$6.26$	$\mathbf{1.810}$	$6.26$	$1.831$	$0$	$0.99$
$90$	$23.26$	$2.107$	$\mathbf{6.53}$	$\mathbf{1.953}$	$2.8$	$1.08$
$450$	$96.61$	$3.066$	$\mathbf{6.90}$	$\mathbf{1.962}$	$14.1$	$1.56$
$4500$	$1506.95$	$26.002$	$\mathbf{11.01}$	$\mathbf{2.303}$	$140.6$	$\mathbf{11.29}$

Table 3: KV-cache inference on RTX 4090: batch 1, 30 active frames, 16 cached text tokens, CUDA Graph, PyTorch SDPA, no CFG. FLOPs and latency include the complete backbone update, causal re-encoding, and K/V writes for the newly finalized frame. Cache counts bf16 history K/V.

Figure 5: Prompt switching along a commanded path. Body colors indicate the active prompt. Blue is the commanded center-of-mass path, red the generated center-of-mass track; numbers show planar speed in m/s. The examples include repeated jumping and backward walking.

Figure 6: Speed control under fixed text prompts. The walking (a) and running (b) sequences follow varying commanded speed profiles. Body and trajectory colors indicate generated planar speed, illustrating gait adaptation along the commanded path while the action prompt remains unchanged.

Appendix figures & tables7 assets

Supplementary material from the paper’s appendix.

Appendix

Figure 7: HumanML3D-calibrated FK matrices. Trace-normalized kinematic-response (263-D) and surface-pullback (138-D) matrices with descending eigenvalue spectra. The 263-D heatmap shows the active 67-channel block; omitted channels have zero weight. Each heatmap uses its own signed color scale.

Panel of Figure 8 — Figure 8: Sampling under three standalone regression losses. (a) Cumulative distributions of the decoded rotation vector projected onto $(1,2,3)^{\top}/\sqrt{14}$ . (b) Probability assigned to each data region, determined by the nearest training rotation in FK output space. Curves and bars show means over two training seeds; whiskers show their ranges. MSE and the FK-induced quadratic retain all eight regions, whereas nonlinear FK concentrates samples in the highest-density region.

Figure 9: Biomechanical center-of-mass mapping to the neutral SMPL-H body. A front torso detail and a full-body side view locate the anatomical landmarks and the mass-weighted whole-body center of mass. The table compares the landmark heights from de Leva (1996) with their SMPL counterparts. The ground root is the planar projection of the resulting center of mass.

SMPL joint	Weight	SMPL joint	Weight
Pelvis	0.05846	Neck	0.07875
L_Hip	0.11777	L_Collar	0.00000
R_Hip	0.11777	R_Collar	0.00000
Spine1	0.05846	Head	0.06940
L_Knee	0.08225	L_Shoulder	0.01146
R_Knee	0.08225	R_Shoulder	0.01146

Appendix

Table 7: SMPL-H node weights for the ground-projected root pivot. Weights follow the segment masses and center locations in de Leva (1996) , using the stated landmark mapping and normalization to unit sum.

Root definition	Jerk $\downarrow$ ( $\mathrm{m/s^{3}}$ )	Curvature $\downarrow$ ( $\mathrm{m^{-1}}$ )
Pelvis projection (263-style)	33.450	32.632
Mass-weighted center-of-mass projection (ours)	26.650	29.504

Appendix

Table 8: Smoothness of the global root definition. Statistics are computed over the complete HumanML3D training split at 30 fps, with each motion weighted equally. Jerk is the mean norm of the third temporal derivative of planar position; curvature is evaluated at frames whose planar speed exceeds

0.05\,\mathrm{m/s}

Representation	Dim.	Pelvis error ( ∘ ) $\downarrow$	Body-local error ( ∘ ) $\downarrow$
HumanML3D ( Guo et al., 2022a )	263	$1.635$	$30.680$
MotionStreamer ( Xiao et al., 2025 )	272	$<10^{-5}$	$<10^{-5}$
FloodDiffusion 2 (ours, 138-D)	138	$<10^{-5}$	$<10^{-5}$

Appendix

Table 9: Source-rotation round trip. Frame-weighted mean geodesic error over the HumanML3D source-motion collection (26,846 motions; two SVD failures excluded for 263-D).

Dataset	Variant	Predicted channels	Checkpoint step	CFG scale
HumanML3D	FD2	263	60K	4
HumanML3D	FD2-Path	260	55K	3
SEED	FD2	138	300K	2
SEED	FD2-Path	135	300K	2

Appendix

Table 10: Dataset-specific checkpoint and inference settings. The number of predicted channels excludes the three clean root channels for FD2-Path.

Explore similar work

Triangular Resampling for Long-Horizon Motion Generation
Sep 28, 2026Kunhang Li, Yiyi Cai, Xiangyue Zhang +5Denoising Trajectory Diffusion Models
The University of Tokyo · Alaya Lab · Japan Advanced Institute of Science and Technology
Stream Forcing: Constructing Unified Training Trajectory for Robust Streaming Video Generation
Aug 11, 2026Yueting Zhu, Yuehao Song, Kaicheng Zhang +5Video Diffusion Models Video Generation
Huazhong University of Science & Technology · Anyverse Dynamics · Horizon Robotics
Reconstruction-Anchored Diffusion Model for Text-to-Motion Generation
Jan 21, 2026Yifei Liu, Changxing Ding, Ling Guo +2Text-To-Motion Generation Text-To-Video Diffusion Models
South China University of Technology · Joy Future Academy

Sep 28, 2026, cs.CVJ/K move · Enter open · S save

Triangular Resampling for Long-Horizon Motion Generation

Kunhang Li, Yiyi Cai, Xiangyue Zhang, Fangyuan Tu+4

The University of Tokyo · Alaya Lab · Japan Advanced Institute of Science and Technology

Denoising Trajectory Diffusion Models Resampling Motion Generation Ground Truth

Open paper AskPDF arXiv

We introduce Triangular Resampling (TR), a post-training method for mitigating long-horizon error accumulation in motion diffusion models. Built on FloodDiffusion's triangular denoising schedule, TR addresses the mismatch between ground-truth-derived training windows and model-generated inference states. Replacing only completed motion history leaves this mismatch unresolved in partially denoised states within the active window. TR therefore extends rollout-based training to these states, using ground-truth clamping to limit excessive drift. For each replayed sample, TR draws one denoising threshold, shared across latent positions and replay updates, and replays multi-step triangular denoising without gradient tracking. After each update, states below the threshold are replaced with noise-matched ground truth, while those at or above it retain model predictions. The resulting latent window enters the standard training update. This rollout construction supports both supervised training (TR) and distribution matching (TR-DMD). On 120-second motion generation from HumanML3D test prompts, TR and TR-DMD achieve state-of-the-art FID AUC within their respective non-DMD and DMD comparison groups. Supervised TR reduces FID AUC by 40.9% and FID degradation slope by 55.3% relative to matched post-training without replay.

Abstract

Explore similar work

Triangular Resampling for Long-Horizon Motion Generation

Stream Forcing: Constructing Unified Training Trajectory for Robust Streaming Video Generation

Reconstruction-Anchored Diffusion Model for Text-to-Motion Generation